Skip to content

Attribution in Slack: what to post and what to skip

Analytics channels get muted for predictable reasons. Four rules about what earns a post, and the single category of update that guarantees people stop reading.

Three platforms, three answers, one budget to defend. Your GA4 export already holds the answer, and the first finding is free and never leaves your browser. Get your first finding free

By , Founder & CEOUpdated 3 min read

Volume trains people to skip, and once trained they skip the one message that mattered.

Four rules

RuleWhy
Post on a fixed cadence, not on changeChange in a causal estimate is expected and not newsworthy
One post, four fields, per channelLong posts get scrolled past regardless of content
Name the widest interval yourselfIt teaches the room to read uncertainty
Never post without a decision promptA number with no question attached is wallpaper

End every post with the actual question: what should we do about the channel with the widest interval, or is anyone against holding this budget steady for a fortnight.

What never to post

Movement alerts. A causal estimate moves week to week because the underlying data is finite and the window shifts. An alert on movement fires constantly, means nothing individually, and is the single fastest way to get a channel muted.

Alert on the pipeline instead: the job failed, the delivery was late, coverage dropped below the floor.

What to post monthly

What changed since last month, and the one decision on the table.

If nothing material moved, say that in a sentence and post nothing else. A channel that is honest about quiet months earns attention in loud ones.

What to post when something genuinely changes

A threshold crossing agreed in advance. Not "Meta went up", but "Meta's interval now sits entirely above the 2.4 threshold we agreed in March, which is the condition we said would trigger a scaling test".

Agreeing thresholds in advance is also what stops the numbers being relitigated afterwards, which is covered in the lift number your agency will accept.

How to build it

The three options are compared in putting attribution results into Slack or Notion.

Before you build

Post it by hand for two cycles and see whether anyone replies. Replies are the signal that the format is right. Silence means fix the format, not the delivery mechanism. The first read costs €99 on a Google Analytics export, refundable within 30 days, no questions asked.

Frequently asked questions

  • Why do analytics Slack channels get muted?
    Because they post things that require no response. Volume trains people to skip, and once trained they skip the message that mattered. Every post should carry a decision prompt, or it is wallpaper regardless of how good the number is.
  • Should I alert when an attribution number changes?
    No. A causal estimate moves week to week because the data window shifts, so a movement alert fires constantly and means nothing individually. Alert on the pipeline instead: job failure, late delivery, or coverage falling below your floor.
  • What makes a threshold alert worth reading?
    That the threshold was agreed before the data arrived. A post saying an interval now sits entirely above a number the team set in March is genuinely informative.

Go deeper: Incrementality testing, explained.

Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.

Have an audience of ecommerce operators? Earn 20% recurring commission by recommending the €99 read.

Keep reading

Terms in this article

Browse the full glossary

Your platforms guess.
We run the math.

Upload a GA4 export and see what each channel caused, next to last-click, in 1–2 minutes. The read is yours to keep.

The free finding runs in your browser. Your file never leaves it. The full read is €99, refundable within 30 days, no questions asked. Prices exclude VAT.